Determinação de um Tamanho Ótimo Amostral para a Análise do Desempenho de Estudantes em Ambientes Educacionais Ubíquos

Abstract

Computational devices have become ubiquitous elements in everyday human life. This current paradigm of interactivity is called Ubiquitous Computing and its usage in academic scope allows the automation of educational activities through the creation and implementation of Ubiquitous Educational Environments. In the literature, several studies analyze these technologies, however they are focused solely on conducting researches to conclude if the usage of new tools leads to improvements in observed variables, not focusing also on initially analyzing if the sample size is sufficient to ensure reliable results. These analyzes may result in inaccurate information, due to the lack of experimental accuracy of the methods used in the studies. Several factors influence experimental precision and reliability, such as heterogeneity of the experimental material under analysis. In order to reduce experimental error, it is essential to use samples with adequate dimensions and that are able to comprehend the observed populations’ characteristics. In this context, this research determined an optimal sample size of 25 students to analyze the performance of students which do not use educational technologies and of 20 students for classes that are in touch with educational systems during academic terms. These results are relevant to researches that make use of human factors, since we demonstrate that samples with reduced amounts of observations are able to comprise most part of the behavior of variables under analysis.